checkstring, clitic bugs/features, vowel length

[helmadik]

I have slowly been finding my way around the system (enjoying it!). Here are some things deep in the bowels that I'd love some help with!
@gregorycrane, maybe?

checkstring:
symptom 1: does not recognize ka/k' as kaka/.
symptom 2: does recognize dei/n' as deina/ (good!)
symptom 3: parses ka/k' as from ka/kh forms. (bad)
symptom 4: parses kak' as kaka/ etc. (bad)
=>

1. elision of non-function words shifts suffix accent to acute on penult. dei/n' ka/k' etc. So not only do we have to add an accent to words like a)ll' (to a)lla/), we also have to consider a shifting accent in the case outlined here.
2. we should check back after evaluating the elision-supplements to see if we have in fact added a short syllable. Any -a- seems to go.
3. symptom 2 suggests there is a fallback in place when no parse is found for the current accentuation. Here and only here I would suggest that we also want to double-check words like ka/k' for possibilities with the acute shifted to ultima: kaka/, kake/,..

clitics:
Trying to add ou=n, dh/pote, pote to words. Added these where -per is in checkstring, but can't seem to convince the system to lose the accent on the preceding syllables. E.g. o(poiondh/pote, o(poia|ou=n o(pw/spote --- Many such forms are currently dealt with individually as wd: which comes out as 'indeclinable', which is kind of sad.

alpha plus iota subscript:
System is happy to treat this as short alpha plus iota.
So, stem-accented neani/a| is treated like neani/ai (the latter also gets the dative parse, but in an age of subscript-avoiders, that is fine..). All problems go away if there is no stem accent or if that vowel preceding the ending is long, e.g. xw/ra| doesn't get the plural reading.

Hipster adscript users
Poor morpheus was written before the adscript revolution. This leads to problems in e.g. the Euripides corpus, with forms like σώιζεται, σώιζει, ἔσωιζον, etc. We need an extra routine (in checkstring?) that, when lost for answers, looks at w/i, w=i, wi not folowed by + (diaeresis), and tries w/|, w=|, w| respectively. I've added single forms for similar things with a)ei/dw, a)/|dw, which we see also as a)/idein :-( PS: I see that forms like a)qw/ious get nicely rewritten to a)qw/\ous and parsed. My suffering is aggravated now.